RichardErkhov/kevin009_-_minirewrite-4bits

Quantization made by Richard Erkhov.

Github

Discord

Request more models

minirewrite - bnb 4bits

• Model creator: https://huggingface.co/kevin009/
• Original model: https://huggingface.co/kevin009/minirewrite/

Original model description:

language: - en license: apache-2.0 tags: - text-generation-inference - transformers - mistral - trl

Model Card: Minimalist Assistant

Model Details

• Base Model: Mistral Instruct v2
• Tokenizer: based on Mistral Instruction following

Intended Use

• As Editor Assistant for revision and paraphrasing
• Avoids technical jargon in favor of clear and accessible language

Training Data

• Initial Training: 14,000 conversations in minimalist style and more accessible language
  • Dataset: kevin009/system-defined-sft-llama3-14k
• Further Training: 8,000 revision conversations to enhance rewriting and paraphrasing tasks.

Performance and Limitations

• Limitations:
  • May produce shorter outputs compared to original version.
  • Potential biases

Ethical Considerations

• Designed for daily use, potential biases from training data should be considered
• The model does not have implemented safety measures to prevent generation of potentially harmful or offensive content

Additional Information

• Fine-tuned to address limitations in writing tasks observed in other models
• Personalized for everyday use cases
• Motivation for development was to create a model better suited for writing tasks, as existing models were found lacking in this area
• SFT fine-tuned model